Jinsook (Jennie) Lee

I study how AI is reshaping the U.S. higher education system through diversity and equity perspectives. I examine (1) how post-SFFA policy has influenced diversity and arbitrariness in machine-driven decisions, and (2) the equity implications of generative AI in admissions, comparing pre- and post-GPT application essays across socio-economic groups. I ask what now counts as an "admissible" identity representation, how AI reshapes it, which populations become vulnerable, and ultimately what human values institutions should look for.

papers The Digital Divide in Generative AI · Ending Affirmative Action Harms Diversity Without Improving Academic Merit

press Inside Higher Ed · Cornell Chronicle, 2024 · Cornell Chronicle, 2025

As part of the early studies at the National Tutoring Observatory, I apply technical approach to improve tutoring move annotation. I challenge the current utterance-level annotation unit and propose a segmentation method using LLMs and other dialogue-segmentation algorithms.

papers Domain-Adapted Retrieval for In-Context Annotation of Pedagogical Dialogue Acts · Codebook-Injected Dialogue Segmentation · AI Annotation Orchestration

support NSF · Gates Foundation · Chan Zuckerberg Initiative

I explore whether we can steer conversational agents to evaluate relational skills that are hard to discover from text-based applications in hiring and workforce contexts. I focus on capturing signals in simulated dialogue situations that require relational skills, and identifying which signals are most predictive of later performance.

support Laude Institute Moonshots